This article appeared in a journal published by Elsevier. The attachedcopy is furnished to the author for internal non-commercial researchand education use, including for instruction at the authors institution

and sharing with colleagues.

Other uses, including reproduction and distribution, or selling orlicensing copies, or posting to personal, institutional or third party

websites are prohibited.

In most cases authors are permitted to post their version of thearticle (e.g. in Word or Tex form) to their personal website orinstitutional repository. Authors requiring further information

regarding Elsevier’s archiving and manuscript policies areencouraged to visit:

http://www.elsevier.com/authorsrights

Author's personal copy

The Practice of Exposure Therapy:Relevance of Cognitive-Behavioral Theory and Extinction Theory

Jonathan S. AbramowitzUniversity of North Carolina at Chapel Hill

Exposure therapy is the most effective psychologicalintervention for people with anxiety disorders. Whilemany therapists learn how to implement exposure tech-niques through clinical training programs or instructionalworkshops, not all of these educational efforts include afocus on the theory underlying this treatment. Theavailability of treatment manuals providing step-by-stepinstructions for how to implement exposure makes it easierfor clinicians to use these techniques with less training thanthey might otherwise receive. This raises questions regard-ing whether it is necessary to understand the theory behindthe use of exposure. This article argues that knowledge ofthe relevant theory is crucial to being able to implementexposure therapy in ways that optimize both short- andlong-term outcome. Specific ways in which theory is relevantto using exposure techniques are discussed.

Keywords: exposure therapy; anxiety; exinction cognitive-behavioraltheory

ACROSS THE MENTAL HEALTH FIELDS there is a greatdeal of inconsistency in how psychological treat-ments are taught to trainees (and to professionals).While most of this training necessarily focuses ontechnique—how to implement the various treat-ment procedures—considerably less attention isoften paid to helping the trainee understand thetheory that forms the basis for these treatment

procedures. This lack of emphasis on theoreticalmodels might be an unfortunate by-product of thefield’s current (and important) emphasis on treat-ment manuals and outcome research. It also mightbe driven by the (similarly important) need torapidly disseminate effective psychological treat-ments. Another reason theory might be less valuedthan technique is that psychological theories can bedifficult to understand, requiring a large timecommitment that some might feel is not essentialto providing effective treatment. Yet this state ofaffairs begs the question of how effective one can bewhen delivering psychological treatments if there isno understanding of the science behind the treat-ments being delivered.In the present article I will argue that in the case

of exposure therapy for pathological anxiety andfear (i.e., anxiety disorders), knowledge of contem-porary cognitive-behavioral models of anxietydisorders and the principles of extinction (i.e., thetype of learning that occurs with exposure) isextremely important in helping patients achieveoptimal short- and long-term outcome. I will beginwith a description of exposure techniques andreviews of contemporary cognitive-behavioralmodels of anxiety disorders and extinction theoryon which the principles of exposure therapy arebased. After a brief review of research supportingthe efficacy of exposure, I will turn to someanecdotes and observations I have made of novicetherapists who did not have sufficient knowledge ofthe relevant theory. I will then discuss severalreasons supporting my contention that at least aworking knowledge of the theoretical frameworkdiscussed in the first part of this article is vital inobtaining optimal short- and long-term successwith exposure.

Available online at www.sciencedirect.com

ScienceDirectBehavior Therapy 44 (2013) 548–558

www.elsevier.com/locate/bt

Address correspondence to Jonathan S. Abramowitz, Ph.D.,Department of Psychology, University of North Carolina at ChapelHill, Campus Box 3270 (Davie Hall), Chapel Hill, NC 27599;e-mail: jabramowitz@unc.edu.0005-7894/44/548-558/$1.00/0© 2013 Association for Behavioral and Cognitive Therapies. Published byElsevier Ltd. All rights reserved.

Author's personal copy

Exposure Therapy as a Treatment forAnxiety Disorders

Exposure therapy is a set of psychological treatmenttechniques (usually considered a form of behavioralor cognitive-behavioral therapy [CBT]) for the typesof pathological fear that are typically observed inpeople with anxiety disorders (although exposure canalso beused to reduce pathological fear that is not partof an anxiety disorder). The techniques all involvehelping the patient engage in repeated and sometimesprolonged confrontation with a stimulus that pro-vokes fear even though it objectively poses no morethan acceptable (i.e., “everyday”) risk. Feared stimulican be alive (e.g., spiders, people with HIV, clowns),inanimate (e.g., toilets, knives, numbers), situational(e.g., driving, darkness, feeling uncertain), cognitive(e.g., “impure” sexual thoughts, memories of trau-matic events, premonitions of untimely accidents), orphysiological (e.g., racing heart, feeling out of breath,a skin blemish). The aim of exposure is to facilitateextinction—reduction in the conditioned anxiety/fearresponse associated with the feared stimulus. Duringexposure, confrontation with the fear-eliciting stimu-lus typically precipitates an observable response,ranging from mild apprehension to intense fear,based on the person’s exaggerated expectation ofdanger—although this initial fear activation is notnecessary for exposure to produce extinction orbeneficial effects on symptoms (e.g., Foa et al.,1983). Over time, this anxious or fearful responsetypically declines naturally—even in the presence ofthe feared stimulus—a process known as habituation.Here again, research indicates that habituation is nota necessary condition for extinction learning to occurduring exposure (e.g., Rowe&Craske, 1998; but seeCraske et al., 2008, for a review).

cognitive-behavioral modelof anxiety

The use of exposure as a treatment for anxiety andfear-based problems follows froma theoretical modelof clinical anxiety implicating dysfunctional beliefs,classical conditioning, and operant conditioning (e.g.,Barlow, 2002). Patients with clinical anxiety prob-lems are characterized by two types of dysfunctionalcognitions: (a) exaggerated estimates of the likelihoodof harm, and (b) exaggerated estimates of the severityof harm. These undue perceptions of threat underlieanxiety responses to the triggers that characterizethe various anxiety disorders (e.g., social stimuli,“contaminated” items, animals, etc.). Over time, fearmight become a conditioned response to such stimuli.In order to reduce or control the conditioned

anxiety (and reduce the perception of threat), peoplewith anxiety disorders resort to safety behaviors—

forms of active and passive avoidance performed toreduce fears of negative consequences and bringabout a sense of security—which are also character-istic of the various disorders (e.g., avoidance inphobias, compulsive rituals in OCD, anxiolyticmedication use in panic, etc.). Safety behaviors,which often reduce anxiety in the short term (andmore rapidly than would naturally occur), have thelong-term effect of preventing the natural extinctionof classically conditioned fear. Moreover, they arenegatively reinforced (operant conditioning) by thereduction in anxiety they engender, thus becominghabitual. From a cognitive-behavioral perspective,safety behaviors maintain the exaggerated threatperceptions and classically conditioned fear re-sponses by (a) fostering premature escape fromanxiety before it naturally extinguishes, and (b)preventing the disconfirmation of the misperceptionsof threat. For example, following the nonoccurrenceof death from a panic episode, a person with panicdisorder will say that the only reason she did not diewas that her benzodiazepine medication kicked inand reduced her heart rate before her extremeanxiety led to a fatal heart attack. Safety behaviorsthus serve as maintenance processes in anxietydisorders; and the fact that they are negativelyreinforced ensures a self-perpetuating vicious cycle.

how does exposure therapy reduceclinical anxiety and fear?

Two empirically derived theoretical models havebeen articulated to explain the effects of exposuretherapy. The earlier of the two is emotionalprocessing theory (EPT), which was first proposedby Rachman (1980), elaborated by Foa and Kozak(1986), and further revised by Foa and McNally(1996). EPT asserts that confrontation with a fearedstimulus during exposure activates a fear structure—a set of propositions about the feared stimulus (e.g., asocial interaction), response (e.g., trembling, sweat-ing), and their meaning (e.g., people will notice and Iwill be embarrassed) that is stored in memory.Activation of the fear structure, along with integra-tion of information that is incompatible with it, isthought to result in the development of a newnonfearstructure that replaces (Foa & Kozak, 1986) orcompetes with (Foa & McNally) the original one.The basis for this corrective learning (i.e., incompat-ible information) is the habituation (i.e., reduction) offear during an exposure trial and between trials (Foa& Kozak) in the absence of any avoidance or safetybehavior. Thus, according to EPT, initial fearactivation, within-session habituation, and between-session habituation are all indicators of successfullearning (and therefore successful exposure therapy).Put another way, EPT assumes that performance

549exposure therapy

Author's personal copy

during exposure is commensurate with learning: fearreduction at the end of an exposure session representsa change in cognitions (e.g., estimates of danger)while continued fear throughout the session does not.Research, however, does not uniformly support

the main tenets of EPT. That is, neither initial fearactivation nor habituation (within or betweenexposure sessions) is a consistent predictor oftherapeutic outcome with exposure. Indeed, per-formance more generally has not been found to be areliable indicator of learning (Bjork & Bjork,2006). Accordingly, a more recent model toaccount for the effects of exposure focuses oninhibitory mechanisms; and this accounts fordiscrepancies between performance during extinc-tion training and post-extinction levels of fear.Within the context of exposure therapy, inhibitorylearning refers to the notion that fear associationsare not removed during extinction, but ratherremain intact as new learning about the fearedstimulus occurs (e.g., Bouton& King, 1983; Craskeet al., 2008). That is, following successful exposure,the feared stimulus is thought to possess twomeanings: the original excitatory (i.e., fear-based)meaning as well as an inhibitory (“safety-based”)meaning. Thus, even if fear subsides followingsuccessful exposure, the original excitatorymeaning isretained and may be recovered under certain circum-stances such as a change in context (i.e., renewal), thepassage of time (i.e., spontaneous recovery), andreacquisition of the original association (Bouton,2002). Accordingly, from this perspective, the aimof exposure therapy is to help patients develop (a) newnonthreat associations, and (b) ways of enhancing theaccessibility of these new associations (relative to theolder threat-associations) in different contexts andover time.One implication of the inhibitory learningmodel is

that the best indicators of the effects of exposuretherapy are posttest or follow-up assessments, whenthe inhibitory learning acquired during exposure willshape how fear is expressed, independent of whetherhabituation occurred during exposure (e.g., Craskeet al., 2008). A related implication is that duringexposure, fear tolerance is more important than fearreduction. This is consistent with research indicatingthat acceptance of negative emotional states reduceslonger-term distress (Eifert & Heffner, 2003) where-as attempts to control, suppress, avoid, or escapefrom negative emotions (i.e., experiential avoidance)are associatedwithmore severe symptoms of anxietydisorders (e.g., Abramowitz, Lackey, & Wheaton,2009; Berman,Wheaton,McGrath,&Abramowitz,2010; Forsyth, Eifert, & Barrios, 2006; Mahaffey,Wheaton, Fabricant, Berman, & Abramowitz,in press). The aim of fostering fear tolerance

also complements the goal of enhancing inhibitorylearning: to the degree fear is tolerated, inhibitoryassociations (e.g., fear is not dangerous) can bemaximally acquired (Arch & Craske, 2011). Ac-cordingly, demonstrating to patients that they cantolerate fear and “act with anxiety” during and afterexposure may be more important in the long runthan ensuring within- and between-session fearreduction (i.e., habituation; Arch & Craske, 2008).

efficacy of exposure

A vast body of treatment outcome studies andmeta-analyses indicates the efficacy of exposure-based therapy—often on its own, but sometimes incombination with other psychological or pharmaco-logical interventions—for problems involving anxi-ety and fear (e.g., Abramowitz, Deacon, &Whiteside, 2011; Olatunji, Cisler, & Deacon,2010). Accordingly, many exposure-based treatmentprotocols have attained the designation of “well-established treatments” in the American Psycholog-ical Association’s review of evidence-basedtreatments (to qualify for this label a treatmentmust have two or more controlled trials [by separateresearch teams] demonstrating its superiority toplacebo; Chambless & Ollendick, 2001). Currently,treatment programs that emphasize exposure thera-py have attained this level of support for thefollowing psychological disorders: panic disorderwith and without agoraphobia, OCD, and specificphobia. In addition, exposure-based treatmentfor PTSD, social anxiety, and childhood anxietydisorders have achieved the level of “probablyefficacious” (i.e., two studies in which the treatmentis more effective than control, or a series of single-case experimental studies). Several other institutions(e.g., the American Psychiatric Association, theNational Institute for Health and Clinical Excellence)have designated exposure-based cognitive-behavioraltreatments as the intervention with the most researchsupport, recommending that they be considered thefirst-line psychological treatment for OCD, PTSD,and panic disorder.

Some Observations and AnecdotesBefore turning to the reasons that the theorydiscussed above is highly relevant and applicablein the everyday practice of exposure therapy, Iwould like to share three observations that I haveoften made in my work training and supervisingnovice clinicians, consulting with other mentalhealth professionals, and giving training workshopson the use of exposure therapy for anxiety. Theseanecdotes illustrate what can occur when theory isnot considered in the conceptualization of anxiouspatients or in the implementation of exposure

550 abramowitz

Author's personal copy

therapy. For me, these observations underscorehow important it is that exposure therapists have asolid grounding in the theory underlying the use ofthese procedures.

observation 1: “just relax”

Perhaps motivated by concerns that the act ofpurposely provoking fear during exposure therapyis (a) not healthy for the patient, (b) makes anxietysymptoms worse, or (c) will lead to prematurediscontinuation of treatment, some therapists insiston teaching relaxation skills to anxious or fearfulpatients to use while confronting the feared stimuli.On the surface, there is intuitive appeal to the ideathat therapy for anxious patients should involvelearning to relax (after all, inducing fear viaexposure seems like the opposite of what a therapyfor anxious people should include). However, forseveral reasons, this is not a good recipe for long-termfear reduction.Indeed, using relaxation during exposure is

inconsistent with the theoretical models and treat-ment outcome results discussed above. Researchconsistently demonstrates that despite its temporarilyanxiety-provoking nature, exposure helps providelong-term relief from most anxiety and fear-basedproblems. Relaxation, on the other hand, is oftenused as a control/placebo intervention in anxietydisorder treatment outcome studies because it is notexpected to work (e.g., Fals-Stewart, Marks, &Schafer, 1993). On a theoretical level, according toEPT, repeated and prolonged exposure worksbecause it allows the patient an opportunity toexperience the activation and natural reduction offear in the presence of feared stimuli. Relaxationwould deny the patient such an experience. From aninhibitory learning perspective, teaching patients touse relaxation during exposure is inconsistent withthe emphasis on tolerating anxiety, as opposed totrying to reduce it. For some patients who even fearthe experience of anxiety itself (e.g., those with panicdisorder; e.g., “Whenmy heart beats rapidly, I worryI will have a heart attack”), exposure helps themconfront this harmless albeit uncomfortable emo-tion. Relaxation, on the other hand,might reduce thephysiological responses to anxiety in the short-term,but it does not provide long-term relief in the form ofnew learning about the feared stimulus to competewith older threat expectancies.Another problem with using relaxation along

with exposure is that relaxation creates a specificcontext in which extinction might occur and safetymight be learned. For example, a patient whoconducts exposure to elevators while also usingrelaxation will learn that she can ride elevators aslong as she is relaxed. But this is not likely to be a

good long-term solution because sooner or laterconfrontation with the fear stimulus outside theextinction context will cause a recovery of the fear(Bouton, 2002). Thus, if the patient is unable toachieve relaxation, she might again experience fearassociated with elevators, increasing her risk ofrelapse. Put another way, preventing renewal of thefear depends (at least in part) on learning that thefear stimulus is safe in many different contexts(including different states of arousal).

observation 2: when exposure is thehammer, everything is a nail

In contrast to some treatment providers who mightshy away from provoking fear during a therapysession, others get carried away with the idea that“facing your fears” might help with overcomingany problem. About 10 years ago, for example, aphysician colleague referred to me a patient withsevere anger problems and insisted I could treat thispatient “the same as if he had OCD—just exposehim to situations that make him angry until hisanger goes away.” As nice as it would be ifconfronting a stressor reduces any type of negativeemotional response, I explained to the psychiatristthat exposure is a treatment for conditioned fearresponses (not anger) that works via extinction andthat it fosters a change in dysfunctional expecta-tions of threat. Anger, a different type of emotionalresponse, is associated with different types of(although equally as strongly held) dysfunctionalcognitions (e.g., personalization, rigidity) that arenot subject to change via extinction learning in theway that threat-based associations are. While it isindeed helpful for people who have learned otheranger-management strategies to practice confront-ing anger-provoking situations and using theirnewly learned strategies, this is not the same asusing exposure to promote the habituation orextinction of angry responses (Novaco, 1975).Because anger does not habituate in the way thatfear does, empirically supported anger-managementprograms do not include extinction-based interven-tions, but rather focus on cognitive therapy andtraining in assertiveness skills (e.g., Moon & Eisler,1983).I have also encountered clinicians attempting to

use exposure to treat problems such as impulsecontrol disorders (e.g., compulsive gambling),depression, and bipolar disorder (manic symp-toms). Yet none of these problems are characterizedby the same types of dysfunctional cognitions orclassical and operant conditioning mechanisms thatare present in anxiety or that lend themselves to theuse of exposure as an intervention. That is, they donot involve overestimates of threat or conditioned

551exposure therapy

Author's personal copy

fear that is maintained by avoidance and safetybehaviors, and therefore are not good targets forexposure. As with anger, neither depression, mania,nor the urge to engage in hair-pulling habituates(or can be extinguished) in the way that fear doesduring exposure (e.g., Abramowitz et al., 2011).In none of these instances would it be advantageousto help the patient develop nonthreat associations tocompete with threat-related expectancies (since thereare no threat associations to begin with). Moreover,there are other treatment approaches (e.g., habitreversal, stimulus control) to which habit andimpulse control disorders respond preferentially.1

observation 3: overconcern with ha-bituation and subjective units ofdistress (suds) levels

A question that exposure therapy trainees and super-visees often ask is, “How low should the patient’sanxiety be before I stop the exposure session?” Yetone need not be overconcerned with the habituationof anxiety during exposure. Although it is a centraltenet of the EPT view, the emphasis on fear reductionduring exposure runs counter to the inhibitorylearning perspective that practicing fear tolerance ismore likely to enhance extinction in the long run. Tobe sure, the anxiety evoked by exposure is unlikely topersist indefinitely or spiral“out of control” and causeharm to the patient. Thus, exposure to prolongedperiods of anxious responding might be as beneficialas exposure to the actual stimulus that evokes thisemotion. Accordingly, exposure therapists who areaware of the theoryunderlying this treatment embracethe importance of their patients accepting feelings offear and anxiety (these feelings are, after all, a normalpart of life). And if that is an important goal ofexposure, then there is no particular SUDS-basedstopping rule for exposure trials.

what do the observations suggest?

I relate these observations (which I imagine arefairly rampant in the vast mental health-care field)to illustrate the gap between theory and clinical

practice, and to show what can happen when aclinician does not have a solid grounding in thetheoretical underpinnings of exposure therapy.These anecdotes all suggest the importance ofknowledge of the relevant theory for optimizingtreatment outcome. But what is it about the theorythat is so important? How does this knowledgeprevent one from ending up as an anecdote in thisarticle? In the remainder of this paper I willarticulate seven reasons it is important for cliniciansusing exposure therapy to be knowledgeable aboutthe theory underlying this approach.

Why Is Knowledge of Theory Important?1. knowledge of theory is necessary forunderstanding anxiety disorders in a waythat leads to the use of exposure therapy

Before implementing exposure therapy, it is neces-sary to develop an understanding of the patient’sproblem, which can be used to plan an effectiveintervention. Although one might use the DSM tocome up with a psychiatric diagnosis, such aclassification is largely descriptive and atheoretical.DSM diagnoses are based mainly on lists of signsand symptoms, as opposed to psychological mech-anisms. Yet exposure therapy is not a treatment forobsessive-compulsive disorder (OCD), panic disor-der, or posttraumatic stress disorder (PTSD) (or anyother disorder) per se; it is an intervention toextinguish fear (e.g., Abramowitz et al., 2011;Barlow et al., 2010). Exposure targets psycholog-ical processes such as exaggerated beliefs aboutthreat (e.g., “dogs are dangerous”); and providesopportunities for patients to challenge these beliefs(e.g., “dogs aren’t as dangerous as I’d thought”)and learn that anxiety itself is not something thatneeds to be resisted or avoided. Accordingly, properimplementation of exposure requires a theoreticaltemplate for identifying and conceptualizing thesigns and symptoms of anxiety disorders, and forunderstanding how these problems are maintained.As an example, consider Sam, who presented to

our clinic complaining of three problems. First, hedescribed excessive anxiety and worry that he mightmistakenly hit someonewith his carwhile driving. Asa result, he was avoiding driving on streets crowdedwith pedestrians. Second, he complained of two“compulsions,” the first being hair pulling thatoccurred whenever he was alone and had somedown time (e.g., when in the bathroom); and thesecond being excessive and unnecessary checking,including turning his car around to check theroadside to be sure he had not injured anyone.Third, Sam complained of panic attacks thatoccurred from “out of the blue,” such as while tryingto fall asleep or watch a movie. As a result, he had

1 This is one of the reasons that I disagree with the formation ofa new diagnostic category of Obsessive-Compulsive and RelatedDisorders as proposed in DSM-5, which includes OCD as well ascompulsive hair pulling and skin picking. For one thing, the lattertwo conditions involve distinct psychological processes from OCD;thus, their inclusion in the same category will likely lead toconfusion regarding psychological treatment (especially for clin-icians not well-versed in theory). Second, removing OCD from theanxiety disorders could imply the need for different treatmentapproaches, although it is very clear that OCD involves the samepsychological mechanisms and responds to exposure in the sameway as do other anxiety disorders.

552 abramowitz

Author's personal copy

begun to experience anticipatory anxiety eachevening and was avoiding movie theatres.Sam’s therapist relied on knowledge of the

cognitive-behavioral model of anxiety disorders toconceptualize Sam’s difficulties. His recurrentthoughts about hitting pedestrians were conceptual-ized as obsessions—harmless intrusive thoughts thathad become preoccupations because of the way Samwas interpreting them as indicating that they werepredictive of actual events. His checking behaviorwas viewed as a safety behavior in response to theobsessions. Although checking provided reassuranceand reduced Sam’s obsessional fear in the short term,it did not provide long-term relief, and actuallyprevented the natural extinction of his fear.Sam’s panic was conceptualized as a fear of

(harmless) anxiety-related body sensations, andassessment indicated he was afraid that breathless-ness and rapid heart rate heralded catastrophicmedical consequences. Sam’s therapist knew thatshe would need to assess further to identifyavoidance or safety behaviors aimed at trying tocontrol panic-related body sensations. Finally,Sam’s hair-pulling was conceptualized as a responseto general stress and anxiety, rather than as anOCD-related compulsive ritual. Theoretical modelsof trichotillomania propose that this problem ismaintained primarily by positive reinforcement(Stanley, Swann, Bowers, & Davis, 1992; althoughnegative reinforcement can also be a factor),whereby the pulling behavior results in gratifica-tion. Understanding the theoretical model ofanxiety (and hair pulling) ensured that Sam’stherapist would collect all of the required informa-tion to derive a treatment plan involving exposureto the necessary fear cues and reduction of theproper safety behaviors.

2. knowledge of theory provides aguide for assessment

Cognitive-behavioral theory also drives the function-al assessment from which the exposure treatmentplan is derived. Functional assessment in this contextrefers to the gathering of detailed patient-specificinformation about the factors that increase thelikelihood that a particular target problem (such asexcessive fear/anxiety) will be exhibited. The param-eters of this assessment are derived directly fromcognitive-behavioral theory. Accordingly, an exhaus-tive list of the situations and stimuli (internal andexternal) that trigger anxiety must be obtained; andknowledge of theory allows the therapist to gatherthe most complete information. Consider a patientwith panic disorder and agoraphobia, for example. Atherapist unaware of the cognitive-behavioral modelof panic (e.g., Clark, 1986) might gather information

about the more obvious external (i.e., agoraphobic)fear cues, but fail to assess for the less apparentinternal (interoceptive; i.e., arousal-related bodysensations) cues that trigger panic attacks. Thus,knowledge of theory is necessary to ensure accurateand thorough information gathering.Knowledge of theory is also critical for assessing

safety behaviors and then for determining whichbehaviors are true safety behaviors (Sam’s check-ing, which was performed to reduce obsessionalanxiety) and which are not (Sam’s hair pulling,which was associated with some gratification andnot performed to minimize specific fears). Finally,theory also informs assessment of the cognitivelinks between the fear cues and safety behaviors. Inother words, why do patients perform safetybehaviors? What is the feared consequence ofconfrontation with a fear cue in the absence of asafety behavior? This information is important forengineering exposure exercises that match with thepatient’s fears and can produce new learning toinhibit existing threat associations. Thus, knowl-edge of the cognitive-behavioral model of anxietyprovides a framework to guide individual function-al assessment and the development of an exposuretreatment plan.2

Theoretical models of exposure also inform thetherapist of potential prognostic variables (i.e., thatmight predict outcome) which should be assessed;such as severe depression, overvalued ideation(e.g., poor insight), and extremes of physiologicalarousal (Foa & Kozak, 1986). Abramowitz et al.(2000), for example, found that severely depressedOCD patients, relative to those with less depres-sion, fared less well with exposure treatment,perhaps due to the very strong emotional reactivitythat depressed patients experience when conduct-ing exposures. From an EPT view, this strongreactivity impedes within-session habituation.From an inhibitory learning view, strong reactivitymight lead to experiential avoidance in whichpatients try too rigidly to reduce (as opposed totolerate) their anxiety during exposure. Similarly,poor insight, from an EPT perspective, is thought tohinder between-session habituation, leading toattenuated response with exposure (e.g., Foa,

2Many novice therapists equate functional assessment withdiagnostic assessment; yet while the former is theoretically drivenand idiographic in its approach to understanding the individual’sanxiety symptoms, diagnostic assessment is a nomothetic approachbased more or less on atheoretical diagnostic criteria (usually theDSM). Although diagnostic assessment and classification mighthave some uses, the functional relations between the specificsituations, stimuli, and responses—over and above a diagnosis—provide a richer account of the problem and are most critical whenplanning and implementing effective exposure therapy.

553exposure therapy

Author's personal copy

Abramowitz, Franklin, & Kozak, 1999). Hereagain, poor insight into the senselessness of one’sfear would be expected to lead to experientialavoidance rather than fear toleration, therebyhampering inhibitory learning.

3. knowledge of theory provides aroad map for the therapist

Understanding the theoretical basis for exposuretherapy is critical for clarifying a treatment plan.More specifically, it allows the clinician to pinpointprecise dysfunctional cognitions, identify stimuli touse in exposure, and recognize safety behaviors thatmust be curtailed in order for progress to be made.Exposure therapy begins with the development of anidiosyncratic version of the cognitive-behavioralmodel as it applies to the patient’s particular anxietyand fear (Abramowitz et al., 2011). The therapistuses theory to understand how the particular fearcues trigger catastrophic thinking relating to fearedoutcomes, and how these exaggerated beliefs aremaintained by safety behaviors and other mainte-nance processes. Figure 1 shows an example for apatient with OCD. His obsessions were triggeredby any stimulus that reminded him of religion,such as seeing a church or hearing certain words(e.g., “God,” “sin”). Because they had becomethreat cues, he had become hypervigilant to themand was therefore experiencing frequent unwantedblasphemous thoughts and images (e.g., images ofJesus on the cross with an erection). Furthermore,he would mistakenly misinterpret these unwantedand recurring thoughts as indicating that he was a“bad Christian” and that he was in for severepunishment from God (e.g., eternal damnation).

This interpretation made him fearful and furtherpreoccupied with the unwanted thoughts, and themore this occurred, the more he engaged in safetybehaviors such as prayer, attempted thoughtsuppression, and seeking reassurance from others,to try to reduce his distress and doubt. These safetybehaviors, however, further fueled erroneous andmaladaptive core beliefs about the importance ofand need to control thoughts. Readers familiarwith the effects of thought suppression will notethat attempting to stop one’s thoughts usually leadsto the thought’s return, perpetuating the viciouscycle (e.g., Abramowitz, Tolin, & Street, 2001).A strong knowledge of theory is therefore necessary

for using exposure to test the patient’s exaggeratedbeliefs about the likelihood and severity of perceivedthreats. Knowledge of the cognitive-behavioral theoryof OCD, for example, is necessary for a therapist torecognize that the patient with religious obsessionsdescribed previously likely has difficulties withtolerating what others experience as acceptable levelsof uncertainty associated with his obsessional doubts.Thus, exposure is best used to help the patientconfront this doubt (e.g., via situational exposure toreligious stimuli that provoke blasphemous imagesand imaginal exposure to the possibility of havingcommitted a sin or having to face the fearedconsequences of such behavior). Absent knowledgeof the role of intolerance of uncertainty in OCD, andexposure theory, a therapist might conceptualize thisproblem improperly (e.g., a religious crisis or the needfor reassurance from a clergy member), leading to aless effective intervention.Even in more straightforward presentations of

anxiety disorders, dysfunctional beliefs can rarely

FIGURE 1 Idiosyncratic model of a patient’s OCD symptoms.

554 abramowitz

Author's personal copy

be fully tested unless patients discontinue theiruse of safety behaviors while performing exposure(i.e., response prevention), lest they think “nothingbad happened during exposure, but that’s because Idid my safety behaviors.” Knowledge of thetheoretical model is crucial for understanding theimportance of response prevention, and for imple-menting this technique in a manner that ismaximally therapeutic (e.g., which safety behaviorsto stop, and when). To illustrate, consider a sociallyanxious patient who had difficulty making smalltalk with her co-workers. Functional assessmenthelped her specify her belief that “They won’taccept me because the things that I’m interested inare silly.” Ordinarily, this patient would spendinordinate amounts of time reading on the Internetabout things that her co-workers were interested injust so that she could discuss more “important”things with them. Exposure therefore entailed thepatient starting conversations about her own“silly” interests (e.g., reality TV shows), withoutstudying up on her colleagues’ more “important”interests (e.g., current events). This helped herdiscover that, contrary to her beliefs, her co-workersdid not treat her any differently, and that she wasacceptable even without excessive preparation.

4. knowledge of theory is important whenproviding a treatment rationale

I assert that not only the therapist, but also thepatient needs to understand the theory behind the useof exposure techniques—although perhaps at asomewhat less technical level—in order to optimallyimplement this treatment. A working knowledge oftheory allows the therapist and patient to workcollaboratively to ensure that the assessment andtreatment planning are comprehensive. It alsoensures that the patient understands the rationalefor engaging in exposure. Understanding thisrationale is positively associated with treatmentoutcome because it provides patients with a clearunderstanding of how exposure therapy weakensanxiety symptoms (e.g., Abramowitz et al., 2002).Patients who understand the theoretical underpin-nings of their problem and its treatment are betterable to properly implement exposure techniques and,for example, understand the differences between theeveryday casual “exposure” to fear cues that manypatients encounter (e.g., being in a social situationthat can’t be avoided) and therapeutic exposuredesigned specifically for fear reduction. In the case ofthe former, the “exposure” usually takes placewithout being planned, is relatively brief (i.e.,terminated by deliberate escape from distress), andthe fear that is evoked is usuallyminimized or resistedwith the use of safety cues and behaviors. Therapeu-

tic exposure, on the other hand, is systematic,prolonged, and repeated in different contexts, andit does not involve subjective resistance to the fearthat is provoked.Patients who understand the theoretical underpin-

nings of their anxiety problem and its treatment arealso better able to recognize subtle elements of theirfear (e.g., covert avoidance, reassurance-seekingbehavior) and attend to these aspects when confront-ing feared stimuli and resisting covert safety behaviorssuch as mental rituals. Finally, understanding thetheoretical model helps the patient to anticipate thecognitive changes that occur via exposure (i.e.,challenges to exaggerated threat estimates) andtherefore get the most out of this technique. In orderfor the therapist to competently socialize the patient tothis theoretical model, he or she must have a workingknowledge of the theory him- or herself.Socializing a patient to the cognitive behavioral

model involves synthesizing the information col-lected during the assessment and, in a transparentand collaborative way, placing the patient’s diffi-culties within the theoretical framework. Doing sohelps the patient understand how exposure therapycan be beneficial (even if confronting feared stimuliseems counterintuitive). In fact, when therapistsreport that their patients are having problemsmaintaining adherence to exposure therapy, it isoften the case that they have spent too little timehelping the patient understand (a) this theoreticaland conceptual framework, (b) the long-term bene-fits of exposure, even if it might seem fear-provokingin the short-term, and (c) the importance of learningto live with acceptable risk as opposed to trying tohave a guarantee of safety. Methods for explainingthis material to patients are described elsewhere(e.g., Abramowitz et al., 2011).

5. knowledge of theory is importantfor optimizing learning during exposurepractice

There are important implications of extinctiontheory for how therapists can optimize exposuretherapy. For example, one way that nonthreatassociations are developed is when expectations aremismatched with reality—that is, when an antici-pated negative outcome does not materialize(Rescorla & Wagner, 1972). This highlights theimportance of determining the ideal frequency andduration of exposure sessions needed to surpass therate atwhich negative outcomeswould be anticipated.For example,with certain phobias (e.g., fear of bees), asingle prolonged exposure session (e.g., 5 hours)mightbe more effective than a series of shorter exposuresequaling the same total length (e.g., 5 one-hoursessions) because the longer exposure session provides

555exposure therapy

Author's personal copy

more time to learn that anticipated negativeoutcomes are unlikely (e.g., bee stings). In otherinstances (e.g., social anxiety), multiple brieferexposures will be optimal because the anticipatedoutcomes would occur more quickly.Extinction theory also predicts that learning will

be enhanced when separate feared stimuli that havebeen individually addressed are later combinedduring exposure trials (i.e., “deepened extinction”;Rescorla, 2006). This concept is exemplified in themethods used to treat OCD when situationalexposure (e.g., changing a baby’s diaper) andimaginal exposure (e.g., to obsessional doubts ofbeing a child molester) are subsequently combined(e.g., thinking about molesting the baby whilechanging her diaper; Abramowitz et al., 2011).Exposure therapy for panic disorder and agora-phobia provides another example; in this instanceinteroceptive exposure to feared arousal-relatedsensations (e.g., lightheadedness) and situationalexposure (e.g., walking alone outside the home)may be combined (e.g., hyperventilating for 90seconds and then going out for a walk unaccom-panied; e.g., Barlow & Craske, 1988).Learning theory (e.g., Lovibond, Davis, &

O’Flaherty, 2000) also provides the basis for theuse of response prevention (i.e., excluding the use ofsafety behaviors and cues) during exposure. Indeed,when the absence of an anticipated negativeoutcome (e.g., fainting) during exposure to a fearstimulus (e.g., grocery stores) is attributed to asafety behavior (e.g., escape or breathing into apaper bag), there is no reason to change thethreat-related association. That is, safety behaviorinterferes with the development of new nonthreatassociations (Craske et al., 2008). Thus, exposuretherapists must be ever mindful of patients’ use ofmore or less subtle safety behaviors and avoidancestrategies to ensure that there is no other explana-tion for the nonoccurrence of feared catastrophesother than that there is a low risk of danger.

6. knowledge of theory is important forpreventing relapse following exposuretherapy

If, as proposed by the inhibitory learning viewof exposure, newly learned nonthreat associationsdo not replace previously learned threat associa-tions, but instead compete with them, thenrelapse—i.e., the return of fear—is always possi-ble. Understanding that extinction depends oncontext-dependent learning rather than “unlearn-ing” can help therapists implement exposure in waysthat reduce the opportunities for relapse (Bouton,2002). For example, long-term maintenance is likelyto be facilitated by conducting exposure in multiple

contexts, and especially in environments where thepatient’s fear is usually (and most problematically)encountered. This helps to build the number ofretrieval cues present to assist with recall of the newlylearned nonthreat associations, especially in situa-tions where the fear might have been learned (Bjork& Bjork, 1992, 2006; Estes, 1955). Implementingexposures in varied contexts (e.g., situations, moodstates, times of day, drug states; e.g., Bouton) is alsoimportant to enhance generalization of extinction tonew contexts.

7. knowledge of theory is importantfor troubleshooting

Finally, knowledge of the theory behind exposuretherapy is critical when troubleshooting a number ofbarriers to success. Although this article does notafford the space to cover the full list of such obstacles,a few common ones deserve comment. Numerousother obstacles often encountered in exposure arediscussed in Abramowitz et al. (2011).

NonadherenceThe most common obstacle to successful exposure isthe patient’s refusal to confront his or her fearedstimuli or resist safety behaviors whether in thesession or during homework. One way to circumventnonadherence is to ensure that patients grasp theconceptual model of anxiety and understand howtheir own symptoms are maintained according to thetheory outlined earlier in this article. Second, therationale for exposure must be very clear so thatpatients understand how engaging in challenging andsometimes frightening exposure taskswill reduce theirfear and anxiety in the long term. These two pointsunderscore the importance of both the therapist andpatient understanding the theory behind exposure.Therapists sometimes are tempted to suspend orpostpone exposures when the patient becomes highlyanxious. In most cases, this is discouraged since it canreinforce avoidance patterns and send a message thatthe task is too dangerous or difficult.

Arguments Over the Risks of ExposureIt is easy to fall into the trap of engaging in debateswith patients who argue that the risks involved withdoing exposure exercises (e.g., touching a toilet) orstopping safety behaviors (e.g., not taking one’s bloodpressure every hour) are too high to take, even intreatment. It seems intuitive to try to use logic toconvince the patient to engage in exposure andresponse prevention. Nevertheless, such debatesusually fail, and worse, they are functionally equiv-alent to the maladaptive reassurance-seeking andoveranalyzing behaviors that many anxious patientsengage in on a regular basis, and which contribute tothe maintenance of the anxiety disorder by increasing

556 abramowitz

Author's personal copy

attention to threat cues (e.g., Clark, 1999). Knowl-edge of the relevant theory would lead a therapistto avoid such debates and instead use means suchas motivational interviewing skills (e.g., Miller &Rollnick, 2013) to help patients see exposure therapyas the best long-term option even though it mightinvolve short-term fear and uncertainty.

Therapist Discomfort With Using ExposureThere is likely a strong correlation between comfortworking within an exposure framework and knowl-edge of the relevant theory.As I alluded to earlier, it isnot surprising that therapists who view exposure asunnecessarily painful for the patient would also beunfamiliar with the cognitive-behavioral model ofanxiety disorders and rationale for exposure. Thus, itis important for therapists to understand that anyanxiety experienced during exposure is not danger-ous, but rather relatively short lived and therapeutic.By experiencing and tolerating anxiety, and itsinevitable reduction at some point, patients gainnew knowledge about situations they believed weredangerous, and about their own ability to managedistress. Finally, contrary to somepopularmyths, it isimportant for therapists to understand that reducingfears by exposure does not cause “symptom substi-tution,” or the emergence of newanxiety symptom totake the place of the old ones.

ConclusionsTreatment manuals, considered by many experts tobe essential to outcome research studies (Foa &Meadows, 1997), are used to promote the stan-dardization of therapy procedures across therapistsand patients. Optimally, manuals should delineatethe essential principles of treatment and provideclinicians with session-by-session procedural guide-lines. Sometimes, however, the use of step-by-stepmanuals leads to taking for granted the theorybehind the intervention. Exposure therapy is a set oftherapeutic techniques that requires knowledge ofthe cognitive-behavioral model of anxiety disor-ders, and of how fear extinction works, in orderto be implemented optimally. This article hasaddressed some of the important reasons this isso, and also highlighted what can happen when thistheory is overlooked. It is hoped that in addition toproviding training in the practical side of imple-menting exposure, workshop leaders and clinicaltraining programs incorporate an overview (at thevery least) of the theoretical underpinnings of thishighly efficacious treatment procedure.

ReferencesAbramowitz, J. S., Deacon, B. J., & Whiteside, S. P. H. (2011).

Exposure therapy for anxiety: Principles and practice. NewYork, NY: Guilford Press.

Abramowitz, J. S., Franklin, M. E., Street, G. P., Kozak, M. J.,& Foa, E. B. (2000). Effects of comorbid depression onresponse to treatment for obsessive-compulsive disorder.Behavior Therapy, 31, 517–528.

Abramowitz, J. S., Franklin, M. E., Zoellner, L. A., &DiBernardo, C. L. (2002). Treatment compliance and outcomeof cognitive-behavioral therapy for obsessive-compulsivedisorder. Behavior Modification, 26, 447–463.

Abramowitz, J. S., Lackey, G., & Wheaton, M. G. (2009).Obsessive-compulsive symptoms: The contribution of ob-sessive beliefs and experiential avoidance. Journal ofAnxiety Disorders, 23, 160–166.

Abramowitz, J. S., Tolin, D. F.,& Street, G. P. (2001). Paradoxicaleffects of thought suppression: A meta-analysis of controlledstudies. Clinical Psychology Review, 21, 683–703.

Arch, J. J., & Craske, M. G. (2008). Acceptance andcommitment therapy and cognitive behavioral therapy foranxiety disorders: Different treatments, similar mechanisms?Clinical Psychology: Science and Practice, 15, 263–279.

Arch, J. J., & Craske, M. G. (2011). Addressing relapse incognitive behavioral therapy for panic disorder: Methodsfor optimizing long-term treatment outcomes. Cognitive andBehavioral Practice, 18, 306–315.

Barlow, D. H. (2002).Anxiety and its disorders: The nature andtreatment of anxiety and panic (2nd ed.). New York, NY:Guilford Press.

Barlow, D. H., & Craske, M. G. (1988). Mastery of youranxiety and panic. Albany, NY: Graywind Publications.

Barlow, D. H., Farchione, T. J., Fairholme, C. P., Ellard, K. K.,Boisseau, C. L., Allen, L. B., &May, J. T. E. (2010). Unifiedprotocol for transdiagnostic treatment of emotionaldisorders: Therapist guide. New York, NY: OxfordUniversity Press.

Berman, N. C., Wheaton, M. G., McGrath, P., & Abramowitz,J. S. (2010). Predicting anxiety: The role of experientialavoidance and anxiety sensitivity. Journal of AnxietyDisorders, 24, 109–114.

Bjork, R. A., & Bjork, E. L. (1992). A new theory of disuse andan old theory of stimulus fluctuation. In A. Healy, S.Kosslyn, & R. Shiffrin (Eds.), From learning processes tocognitive processes: Essays in honor of William K. Estes(pp. 35–67). Hillsdale, NJ: Erlbaum.

Bjork, R. A., & Bjork, E. L. (2006). Optimizing treatment andinstruction: Implications of a new theory of disuse. In L.-G.Nilsson & N. Ohta (Eds.),Memory and society: Psychologicalperspectives (pp. 116–140). New York, NY: Psychology Press.

Bouton, M. E. (2002). Context, ambiguity, and unlearning:Sources of relapse after behavioral extinction. BiologicalPsychiatry, 52, 976–986.

Bouton, M. E., & King, D. A. (1983). Contextual control of theextinction of conditioned fear: tests for the associative valueof the context. Journal of Experimental psychology: AnimalBehavior processes, 9, 248–265.

Chambless, D. L., & Ollendick, T. H. (2001). Empiricallysupported psychological interventions: Controversies andEvidence. Annual Review of Psychology, 52, 685–716.

Clark, D. M. (1986). A cognitive approach to panic. BehaviourResearch and Therapy, 24, 461–470.

Clark, D. M. (1999). Anxiety disorders: Why they persist andhow to treat them. Behavior Research and Therapy,37(Suppl 1), S5–S27.

Craske,M., G., Kitcanski, K., Zelokowsky,M.,Mystkowski, J.,Chowdhury, N., & Baker, A. (2008). Optimizing inhibitorylearning during exposure therapy. Behavior Research andTherapy, 46, 5–27.

Eifert, G. H., & Heffner, M. (2003). The effects of acceptanceversus control contexts on avoidance of panic-related

557exposure therapy

Author's personal copy

symptoms. Journal of Behavior Therapy and ExperimentalPsychiatry, 34, 293–312.

Estes, W. K. (1955). Statistical theory of distributional phenom-ena in learning. Psychological Review, 62, 369–377.

Fals-Stewart, W., Marks, A. P., & Schafer, J. (1993). Acomparison of behavior group therapy and individualbehavior therapy in treating obsessive-compulsive disorder.Journal of Nervous and Mental Disease, 181, 189–193.

Foa, E. B., Abramowitz, J. S., Franklin, M. E., & Kozak, M. J.(1999). Feared consequences, fixity of belief, and treatmentoutcome in patients with obsessive-compulsive disorder.Behavior Therapy, 30, 717–724.

Foa, E. B., Grayson, J. B., Steketee, G. S., Doppelt, H. G., Turner,R. M., & Latimer, P. R. (1983). Success and failure in thebehavioral treatment of obsessive-compulsives. Journal ofConsulting and Clinical Psychology, 51(2), 287–297.

Foa, E. B., &Kozak,M. J. (1986). Emotional processing of fear:Exposure to corrective information. Psychological Bulletin,99, 20–35.

Foa, E. B., & McNally, R. J. (1996). Mechanisms of change inexposure therapy. In M. Rapee (Ed.), Current controversiesin the anxiety disorders (pp. 329–343). New York, NY: TheGuilford Press.

Foa, E. B., & Meadows, E. A. (1997). Psychosocial treatmentsfor posttraumatic stress disorder: A critical review. AnnualReview of Psychology, 48, 449–480.

Forsyth, J. P., Eifert,G.H.,&Barrios,V. (2006). Fear conditioningin an emotion regulation context: A fresh perspective on theorigins of anxiety disorders. In M. G. Craske, D. Hermans, &D. Vansteenwegen (Eds.), Fear and learning: From basicprocesses to clinical implications (pp. 133–153). Washington,DC: American Psychological Association.

Lovibond, P. F., Davis, N. R., & O’Flaherty, A. S. (2000).Protection from extinction in human fear conditioning.Behaviour Research and Therapy, 38, 967–983.

Mahaffey, B., Wheaton, M. G., Fabricant, L. E., Berman, N. C.,& Abramowitz, J. S. (in press). The contribution of

experiential avoidance and social cognitions in the predictionof social anxiety. Behavioural and Cognitive Psychotherapy.

Miller, W. R., & Rollnick, S. (2013). Motivational interviewing:Helping people change (3rd ed.). New York, NY: Guilford.

Moon, J. R., & Eisler, R. M. (1983). Anger control: Anexperimental comparison of three behavioral treatments.Behavior Therapy, 14, 493–505.

Novaco, R. (1975). Anger control: The development andevaluation of an experimental treatment. Lexington, MA:D.C. Health.

Olatunji, B. O., Cisler, J., & Deacon, B. J. (2010). Efficacy ofcognitive behavioral therapy for anxiety disorders: A reviewof meta-analytic findings. Psychiatric Clinics of NorthAmerica, 33, 557–577.

Rachman, S. (1980). Emotional processing. Behaviour Researchand Therapy, 18, 51–60.

Rescorla, R. A. (2006). Deepened extinction from compoundstimulus presentation. Journal of Experimental Psychology:Animal Behavior Processes, 32, 135–144.

Rescorla, R. A., &Wagner, A. R. (1972). A theory of Pavlovianconditioning: Variations in the effectiveness of reinforce-ment and nonreinforcement. In A. H. Black, & W. F.Prokasy (Eds.), Classical conditioning II: Current researchand theory (pp. 64–99). New York, NY: Appleton-Century-Crofts.

Rowe, M. K., & Craske, M. G. (1998). Effects of varied-stimulus exposure training on fear reduction and return offear. Behaviour Research and Therapy, 36, 719–734.

Stanley, M., Swann, A., Bowers, T., & Davis, M. (1992). Acomparison of clinical features in trichotillomania andobsessive-compulsive disorder. Behaviour Research andTherapy, 30, 39–44.

RECEIVED: September 16, 2012ACCEPTED: March 1, 2013Available online 13 March 2013

558 abramowitz